ce7cb702218693d08ea6d83464d2457c9250027c

services/vertex-batch/k8s/deployment.yaml

@@ -0,0 +1,77 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: vertex-batch
+  namespace: agentgateway-system
+spec:
+  replicas: 2
+  selector:
+    matchLabels:
+      app.kubernetes.io/name: vertex-batch
+  template:
+    metadata:
+      labels:
+        app.kubernetes.io/name: vertex-batch
+        app.kubernetes.io/instance: vertex-batch
+    spec:
+      terminationGracePeriodSeconds: 75
+      automountServiceAccountToken: false
+      imagePullSecrets:
+        - name: quay-cluster-pull
+      securityContext:
+        runAsNonRoot: true
+        runAsUser: 65532
+        runAsGroup: 65532
+        seccompProfile:
+          type: RuntimeDefault
+      containers:
+        - name: batch
+          image: registry.quorate.internal/quorate/vertex-batch@sha256:a1416b98cc57ae27dde41df27ad696c933d6bad6906f241bc698fc48cf4c7a35
+          args:
+            - -gateway-url=http://agentgateway-proxy.agentgateway-system.svc.cluster.local/v1/chat/completions
+            - -project=ms2k-personal
+            - -bucket=ms2k-personal-vertex-batch
+            - -credentials=/credentials/credentials.json
+          ports:
+            - name: http
+              containerPort: 8080
+          readinessProbe:
+            httpGet:
+              path: /healthz
+              port: http
+          lifecycle:
+            preStop:
+              sleep:
+                seconds: 5
+          resources:
+            requests:
+              cpu: 10m
+              memory: 32Mi
+            limits:
+              memory: 128Mi
+          securityContext:
+            allowPrivilegeEscalation: false
+            readOnlyRootFilesystem: true
+            capabilities:
+              drop: [ALL]
+          volumeMounts:
+            - name: credentials
+              mountPath: /credentials
+              readOnly: true
+      volumes:
+        - name: credentials
+          secret:
+            secretName: agentgateway-vertex-ai
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: vertex-batch
+  namespace: agentgateway-system
+spec:
+  selector:
+    app.kubernetes.io/name: vertex-batch
+  ports:
+    - name: http
+      port: 8080
+      targetPort: http

services/vertex-batch/k8s/kustomization.yaml

@@ -0,0 +1,6 @@
+apiVersion: kustomize.config.k8s.io/v1beta1
+kind: Kustomization
+resources:
+  - deployment.yaml
+  - routes.yaml
+  - network-policy.yaml

services/vertex-batch/k8s/network-policy.yaml

@@ -0,0 +1,91 @@
+apiVersion: cilium.io/v2
+kind: CiliumNetworkPolicy
+metadata:
+  name: vertex-batch
+  namespace: agentgateway-system
+spec:
+  endpointSelector:
+    matchLabels:
+      app.kubernetes.io/name: vertex-batch
+  ingress:
+    - fromEntities: [host]
+      toPorts:
+        - ports:
+            - port: "8080"
+              protocol: TCP
+    - fromEndpoints:
+        - matchLabels:
+            gateway.networking.k8s.io/gateway-name: agentgateway-proxy
+      toPorts:
+        - ports:
+            - port: "8080"
+              protocol: TCP
+  egress:
+    - toEndpoints:
+        - matchLabels:
+            gateway.networking.k8s.io/gateway-name: agentgateway-proxy
+      toPorts:
+        - ports:
+            - port: "80"
+              protocol: TCP
+    - toEndpoints:
+        - matchLabels:
+            k8s:io.kubernetes.pod.namespace: kube-system
+            k8s-app: kube-dns
+      toPorts:
+        - ports:
+            - port: "53"
+              protocol: UDP
+            - port: "53"
+              protocol: TCP
+    - toEntities: [world]
+      toPorts:
+        - serverNames:
+            - oauth2.googleapis.com
+            - aiplatform.googleapis.com
+            - storage.googleapis.com
+          ports:
+            - port: "443"
+              protocol: TCP
+---
+apiVersion: cilium.io/v2
+kind: CiliumNetworkPolicy
+metadata:
+  name: batch-gateway
+  namespace: agentgateway-system
+spec:
+  endpointSelector:
+    matchLabels:
+      gateway.networking.k8s.io/gateway-name: agentgateway-proxy
+  ingress:
+    - fromEndpoints:
+        - matchLabels:
+            app.kubernetes.io/name: vertex-batch
+      toPorts:
+        - ports:
+            - port: "80"
+              protocol: TCP
+  egress:
+    - toEndpoints:
+        - matchLabels:
+            app.kubernetes.io/name: vertex-batch
+      toPorts:
+        - ports:
+            - port: "8080"
+              protocol: TCP
+---
+apiVersion: cilium.io/v2
+kind: CiliumClusterwideNetworkPolicy
+metadata:
+  name: host-vertex-batch
+spec:
+  nodeSelector: {}
+  egress:
+    - toEndpoints:
+        - matchLabels:
+            k8s:io.kubernetes.pod.namespace: agentgateway-system
+            app.kubernetes.io/name: vertex-batch
+      toPorts:
+        - ports:
+            - port: "8080"
+              protocol: TCP

services/vertex-batch/k8s/routes.yaml

@@ -0,0 +1,88 @@
+apiVersion: agentgateway.dev/v1alpha1
+kind: AgentgatewayBackend
+metadata:
+  name: vertex-batch
+  namespace: agentgateway-system
+spec:
+  ai:
+    provider:
+      vertexai:
+        projectId: ms2k-personal
+        region: global
+      host: vertex-batch.agentgateway-system.svc.cluster.local
+      port: 8080
+---
+apiVersion: gateway.networking.k8s.io/v1
+kind: HTTPRoute
+metadata:
+  name: batch-front
+  namespace: agentgateway-system
+spec:
+  parentRefs:
+    - name: agentgateway-proxy
+  rules:
+    - matches:
+        - method: POST
+          path:
+            type: Exact
+            value: /v1/chat/completions
+          headers:
+            - name: x-model
+              type: RegularExpression
+              value: ^vertex_ai_batch/.+
+      filters:
+        - type: URLRewrite
+          urlRewrite:
+            path:
+              type: ReplaceFullPath
+              replaceFullPath: /front
+      backendRefs:
+        - name: vertex-batch
+          port: 8080
+---
+apiVersion: gateway.networking.k8s.io/v1
+kind: HTTPRoute
+metadata:
+  name: batch-native
+  namespace: agentgateway-system
+spec:
+  parentRefs:
+    - name: agentgateway-proxy
+  rules:
+    - matches:
+        - method: POST
+          path:
+            type: Exact
+            value: /v1/chat/completions
+          headers:
+            - name: x-model
+              type: RegularExpression
+              value: ^vertex_batch/.+
+      backendRefs:
+        - group: agentgateway.dev
+          kind: AgentgatewayBackend
+          name: vertex-batch
+---
+apiVersion: agentgateway.dev/v1alpha1
+kind: AgentgatewayPolicy
+metadata:
+  name: batch-native
+  namespace: agentgateway-system
+spec:
+  targetRefs:
+    - group: gateway.networking.k8s.io
+      kind: HTTPRoute
+      name: batch-native
+  traffic:
+    transformation:
+      request:
+        set:
+          - name: x-batch-stream
+            value: 'string(has(json(request.body).stream) ? json(request.body).stream : false)'
+          - name: x-batch-model
+            value: 'json(request.body).model.stripPrefix("vertex_batch/")'
+  backend:
+    ai:
+      transformations:
+        - field: model
+          expression: llmRequest.model.stripPrefix("vertex_batch/")